THE ARTS
CHILD POLICY
This PDF document was made available from www.rand.org as a public
service of the RAND Corporation.
CIVIL JUSTICE
EDUCATION
ENERGY AND ENVIRONMENT
Jump down to document6
HEALTH AND HEALTH CARE
INTERNATIONAL AFFAIRS
NATIONAL SECURITY
POPULATION AND AGING
PUBLIC SAFETY
SCIENCE AND TECHNOLOGY
SUBSTANCE ABUSE
The RAND Corporation is a nonprofit research
organization providing objective analysis and effective
solutions that address the challenges facing the public
and private sectors around the world.
TERRORISM AND
HOMELAND SECURITY
TRANSPORTATION AND
INFRASTRUCTURE
WORKFORCE AND WORKPLACE
Support RAND
Purchase this document
Browse Books & Publications
Make a charitable contribution
For More Information
Visit RAND at www.rand.org
Explore RAND National Defense Research Institute
View document details
Limited Electronic Distribution Rights
This document and trademark(s) contained herein are protected by law as indicated in a notice appearing
later in this work. This electronic representation of RAND intellectual property is provided for noncommercial use only. Permission is required from RAND to reproduce, or reuse in another form, any
of our research documents for commercial use.
This product is part of the RAND Corporation documented briefing series. RAND
documented briefings are based on research briefed to a client, sponsor, or targeted audience and provide additional information on a specific topic. Although documented
briefings have been peer reviewed, they are not expected to be comprehensive and may
present preliminary findings.
What Factors Affect
the Military Enlistment
of Hispanic Youth?
A Look at Enlistment Qualifications
Beth J. Asch, Christopher Buck, Jacob Alex Klerman,
Meredith Kleykamp, David S. Loughran
Prepared for the Office of the Secretary of Defense
Approved for public release; distribution unlimited
The research described in this report was prepared for the Office of the Secretary of Defense
(OSD). The research was conducted in the RAND National Defense Research Institute, a
federally funded research and development center sponsored by the Office of the Secretary
of Defense, the Joint Staff, the Unified Combatant Commands, the Department of the
Navy, the Marine Corps, the defense agencies, and the defense Intelligence Community
under Contract DASW01-01-C-0004.
ISBN 0-8330-3839-7
The RAND Corporation is a nonprofit research organization providing objective analysis
and effective solutions that address the challenges facing the public and private sectors
around the world. RAND’s publications do not necessarily reflect the opinions of its
research clients and sponsors.
R® is a registered trademark.
© Copyright 2005 RAND Corporation
All rights reserved. No part of this book may be reproduced in any form by any electronic or
mechanical means (including photocopying, recording, or information storage and retrieval)
without permission in writing from RAND.
Published 2005 by the RAND Corporation
1776 Main Street, P.O. Box 2138, Santa Monica, CA 90407-2138
1200 South Hayes Street, Arlington, VA 22202-5050
201 North Craig Street, Suite 202, Pittsburgh, PA 15213-1516
RAND URL: http://www.rand.org/
To order RAND documents or to obtain additional information, contact
Distribution Services: Telephone: (310) 451-7002;
Fax: (310) 451-6915; Email: order@rand.org
Preface
An ongoing concern of Congress, the U.S. Department of Defense (DoD), and the armed
services is the social representation of the military. As there is virtually no lateral entry, a key
concern in particular is the social representation of active duty accessions. Hispanics are a
growing segment of the youth population yet have been underrepresented historically among
military accessions. The widely cited reason is their below-average rate of graduation from
high school and the services’ preference for diploma graduate recruits. But other less-studied
factors may contribute. These other possible factors may include lack of language proficiency
as reflected in their applicant aptitude test scores, fertility choices, health factors such as obesity, and involvement in risky activities such as the use of illegal drugs. These factors could
adversely affect the ability of some Hispanic youth to meet service enlistment standards.
Our project “Hispanic Youth in the United States and the Factors Affecting Their
Enlistment” is conducting an analysis of the factors that lead to the underrepresentation of
Hispanic youth in the military. This documented briefing provides a summary of interim
results from the first year of the research effort. First, it highlights the various enlistment
standards recruits must meet in each branch of service. Then using available data on American young adults, it summarizes the degree to which Hispanic youth qualify for service relative to youth who are members of other racial and ethnic groups. A more detailed summary
of the first-year analysis, together with the summary of the results of the second year, will be
published in a final report for the project. This documented briefing is intended for individuals interested in military recruiting and in the population representation of personnel in
the armed forces.
This research was sponsored by the Defense Human Resource Activity and the Office of Accession Policy, Office of the Under Secretary of Defense for Personnel and Readiness, and conducted within the Forces and Policy Resources Center of the RAND National
Defense Research Institute, a federally funded research and development center sponsored by
the Office of the Secretary of Defense, the Joint Staff, the Unified Combatant Commands,
the Department of the Navy, the Marine Corps, the defense agencies, and the defense Intelligence Community.
Comments are welcome and may be addressed to Beth Asch at Beth_Asch@rand.org.
For more information on RAND’s Forces and Policy Resources Center, contact the Director,
James Hosek. He can be reached by email at James_Hosek@rand.org; by phone at 310-3930411, extension 7183; or by mail at the RAND Corporation, 1776 Main Street, Post Office
Box 2138, Santa Monica, California 90407-2138. More information about RAND is available at www.rand.org.
iii
Contents
Preface .................................................................................................iii
Table .................................................................................................. vii
CHAPTER ONE
Introduction ........................................................................................... 1
CHAPTER TWO
Year One, Task One: Cataloging Current Enlistment Standards .................................. 9
CHAPTER THREE
Year One, Task Two: Assessing the Importance of Enlistment Standard Characteristics
Among Hispanics ................................................................................ 19
CHAPTER FOUR
Year One, Task Three: Examining How the Fraction Passing Enlistment Standards Has
Changed Over Time ............................................................................. 47
CHAPTER FIVE
Year Two: Remaining Research Tasks .............................................................. 57
APPENDIX A
Revising Estimates of Hispanic Representation in the 18- to 24-Year-Old Population ......... 59
APPENDIX B
Enlistment Documents by Source ................................................................... 63
APPENDIX C
White and Black Males and Black, Hispanic, and White Females ................................ 65
References............................................................................................. 71
v
Table
A.1. Share of Hispanics Among 18- to 24-Year-Olds ................................................ 61
vii
CHAPTER ONE
Introduction
1
2
What Factors Affect the Military Enlistment of Hispanic Youth?
Hispanic Accessions Have Grown, but Are
Not Same Share as Population
Percent Hispanic,
18- to 24-Year-Old Civilians, DoD Accessions
18
Percent
16
14
12
10
8
6
1994
1996
1998
2000
2002
2004
Hispanic Civilians, Pop Rep
DoD Hispanic Accessions
SOURCE: DoD (2004) (Pop Rep).
-2
Although the representation of Hispanics among military recruits has increased in
the past decade, this group is still underrepresented among military enlistments. This chart
shows the fraction of Department of Defense (DoD) accessions that is Hispanic and the fraction of 18- to 24-year-old civilians in the U.S. population who are Hispanic since 1994. The
Department of Defense tracks information about the demographic composition of its personnel and of the 18- to 24-year-old civilian population in an annual report, Population Representation in the Military Services (Pop Rep). Pop Rep’s estimates show an increase in the fraction of the youth population that is Hispanic (the upper line in the figure). According to Pop
Rep, the fraction of the youth population that is Hispanic increased by about 20 percent
from 1994 to 2002 (from 13.0 percent to 15.6 percent).
The fraction of DoD enlisted accessions who are Hispanic has also increased, according to the Pop Rep report, from 7.6 percent in 1993 to 11.3 percent in 2002. Thus military accessions seem to be catching up but are still underrepresented in terms of Hispanic
recruits.
Underrepresentation is a concern to the Congress, DoD, and the services. As stated
in Pop Rep, “There is a continuing need to track demographic changes and to monitor the
balance between benefits and burdens across the varied segments of society” (DoD, 2004,
pg. ix).
Introduction
3
An Aside: Use of 1990 Census Weights in
Pop Rep Undercounts Hispanic Population
Percent Hispanic of 18- to 24-Year-Old Civilians,
Using 1990 or 2000 Census Weights
20
Percent
18
16
14
12
10
1994
1996
1998
Hispanic, Pop Rep
Hispanic, 2000 Weights
2000
2002
2004
Hispanic, 1990 Weights
SOURCES: DoD (2004); U.S. Census Bureau (1990, 2000).
-3
As an important aside, it seems clear that Pop Rep’s estimates of the fraction of the
youth population that is Hispanic are too low. As we discuss in Appendix A, the results of
the 2000 U.S. Census suggest that the 1990 Census significantly under-counted Hispanics.
However, through its 2002 edition, Pop Rep has used official Census Bureau estimates or
weights that are based, in part, on the 1990 Census results. Once Pop Rep switches to new
Census Bureau weights based on the 2000 Census, its estimates of the share of the population that is Hispanic will jump sharply.
The 2000 Census-based weights are preferred. The upper line in the above chart uses
2000 Census-based weights to estimate the share of Hispanics in the youth population from
1994 to 2004. These estimates imply that the true fraction of the youth population that is
Hispanic is about two percentage points higher than the official estimates in Pop Rep (e.g.,
17.1 percent versus 14.9 percent in 2000).
4
What Factors Affect the Military Enlistment of Hispanic Youth?
Representation Varies by Service
Percent of Non–Prior-Service Accessions
Who Are Hispanic with Civilian
Comparison, FY 2004
20
Percent
15
10
5
0
Army
Navy
Marine
Corps
Air Force
Total DoD
18- to 24Year-Old
Civilians
SOURCE: Accessions: DoD (undated); Civilians: U.S. Census Bureau (2000, 2004).
-4
The degree of underrepresentation varies by service. According to Pop Rep, the Air
Force has been the least representative of Hispanics, with only 7.3 percent of accessions being Hispanic compared with 15.7 percent of 18- to 24-year-olds in 2002. The Marine Corps
is the closest to being representative of Hispanic youth.
The chart shows more recent representation figures for FY 2004, available from the
Defense Manpower Data Center’s (DMDC’s) Recruit Market Information System (DoD,
undated), and shows more progress by service.1 For example, the system reports the percentage of total accessions that are Hispanic and Latino for the Army, Navy, Air Force, and Marine Corps as 12.5 percent, 15.1 percent, 9.5 percent, and 16.1 percent, respectively, resulting in a DoD figure of 15.1 percent. In comparison, Hispanics represent 17.9 percent of the
18-to-24 population of civilians in 2004. These figures are not directly comparable to pre2003 percentages because of governmentwide changes in how race and ethnicity are defined
in federal data collection efforts. Still, they clearly show progress when compared to the earlier years.
____________
1
One wrinkle in comparing social representation over time is that the definition of racial and ethnic representation is not
comparable to the earlier years because of a governmentwide change in the standard definitions of race and ethnicity in
federal data collections as of January 1, 2003. Among other things, agencies, including the Department of Defense, must
offer individuals the opportunity to select one or more races when reporting race and the categories for ethnicity must include “Hispanic or Latino” and “Not Hispanic or Latino.”
Introduction
5
Underrepresentation Occurs Despite
Higher Reported Enlistment Propensity
Percent
Youth Composite Active Duty Propensity: Percent
Likely or Very Likely to Join a Military Service, Youth
Poll Spring 2004, Ages 16–21
50
45
40
35
30
25
20
15
10
5
0
45
39
31
24
20
10
White
Males
Black
Males
Hispanic White
Black Hispanic
Males Females Females Females
SOURCE: Defense Human Resources Activity (2004, Table 2.2).
-5
The observed demographic composition of new accessions may simply reflect low
propensity among Hispanics. The reality is just the opposite. Surveys of youth attitudes toward the military consistently find Hispanics have a higher propensity to enlist than whites.
The responses from a spring 2004 poll of American youth conducted by the Department of Defense are illustrative. This chart summarizes the responses of youth, ages 16
to 21, to the question, “How likely is it that you will be serving on active duty in (random
service named) in the next few years?” The chart shows the fraction who responded that they
were definitely or probably likely to serve. Among males, Hispanics had the highest percent
responding probably likely or very likely to join relative to white or black males. Among females, 31 percent of Hispanic youth stated they were very likely or probably likely to serve in
the military, significantly more than white females.
6
What Factors Affect the Military Enlistment of Hispanic Youth?
Other Factors May Affect
Underrepresentation of Hispanic Enlistments
Traditional Explanation:
• Other Explanations?
Percent High School Dropouts,
Ages 16–24, 2001
35
30
25
20
15
10
NonHispanic
White
NonHispanic
Black
Hispanic
– AFQT, fertility choices, health
factors, risky behavior
– Enlistment processes
including waivers
– Career and educational
choices made by qualified
Hispanic youth
– Hispanic population is diverse
and factors may vary among
subpopulations
5
0
SOURCE: National Center for Education Statistics (2002)
using Current Population Survey data.
-6
The traditional explanation is Hispanics’ higher than average high school dropout
rate. The services strongly prefer the enlistment of recruits with a high school diploma, aiming to achieve at least a 90-percent annual target. Among individuals in 2001, ages 16 to 24
who are not enrolled, 31.6 percent of Hispanics were high school dropouts. As will be shown
later, high school graduation is an important explanatory variable.
Still, other factors may also be important in explaining both low enlistments and low
high school graduation rates among Hispanic youth. The services have numerous eligibility
criteria in addition to high school diploma status, including language proficiency and aptitude as measured by the Armed Forces Qualification Test (AFQT), and these criteria may
disproportionately disqualify Hispanic youth. Furthermore, how the services implement the
standards, with the ability of the services to waive standards in some circumstances, may affect the representation of Hispanics in accessions. In addition, qualified Hispanic youth may
not be likely to join the military. Therefore it would be useful to investigate what educational
and work choices qualified Hispanic youth make as they transition from high school to stable employment. Finally, given the diversity of the Hispanic population in terms of country
of origin and geographic region with the United States, factors affecting eligibility to enlist
may vary by subpopulation.
Introduction
7
Research Questions
• What are the most significant obstacles to the
enlistment of Hispanic youth, and how have these
varied over time?
• Are there differences in enlistment rates by Hispanic
subgroup?
• What career and educational choices do Hispanic
youth make that affect their enlistment decisions?
• How do the military’s recruiting and waiver processes
affect Hispanic enlistment representation and the
subsequent success of Hispanic recruits?
-7
This research project addresses a series of questions about the Hispanic youth population and their enlistment behavior. First we analyze which enlistment factors are the most
difficult for Hispanics to meet, whether the factors have changed over time, and how they
vary among Hispanic subpopulations to the extent we can identify them. We also consider
the question of the educational and career choices made by Hispanic youth and seek to understand how waivers and other service recruiting practices determine the representation of
Hispanic youth among enlistments. The project also considers the question of how such
practices affect subsequent success in service for Hispanic recruits, measured in terms of metrics such as first-term attrition rates.
CHAPTER TWO
Year One, Task One: Cataloging Current Enlistment Standards
Research Tasks
Year 1
• Task 1: Catalog current enlistment standards
• Task 2: Assess the importance of enlistment standard
characteristics among Hispanics
• Task 3: Examine how the fraction passing enlistment
standards has changed over time
Year 2
• Task 4: Estimate the representation of Hispanics in
military accession by Hispanic subgroup
• Task 5: Analyze the career and schooling paths of
Hispanics
• Task 6: Analyze MEPCOM attrition among Hispanic
applicants
• Task 7: Analyze the waiver process and the effects of
enlistment standards on the success of Hispanic recruits
during their first enlistment term
-8
The project is a two-year project with seven tasks. This documented briefing
summarizes the results of the first year, which covers the first three tasks. These first three
tasks focus on identifying the characteristics that define whether individuals are qualified to
enlist and on analyzing the degree to which the Hispanic youth population exhibit those
characteristics. Thus, the first-year tasks characterize the pool of qualified Hispanic youth
and examine what fraction of the Hispanic youth population would meet the various
enlistment standards.
The remaining four tasks, being conducted in the second year, focus on providing
more refined estimates of the representation of Hispanics among recent and current enlistees,
examining the career and school choices of Hispanic young adults, analyzing Military
Entrance Process Command (MEPCOM) attrition among Hispanic military applicants, and
analyzing the services’ waiver processes and their effects on the personnel outcomes of
Hispanic recruits relative to the recruits.
9
10
What Factors Affect the Military Enlistment of Hispanic Youth?
For the remainder of the briefing, we will first briefly review the services’ enlistment
standards. Next, we will discuss our analysis of the enlistment standard characteristics among
Hispanic youth and discuss how they have changed over time.
Year One, Task One: Cataloging Current Enlistment Standards
11
Enlistment Standards Fall into
12 Broad Categories
1. Citizenship
7. Moral standards
2. English language proficiency 8. Financial restrictions
3. Weight/height
9. Homosexual conduct
4. Education
10. Age
5. Aptitude
11. Strength
6. Medical
12. Dependents
-9
To analyze what fraction of Hispanic youth would meet the various enlistment
standards, we first needed to catalogue the enlistment standards. Enlistment standards are
highly detailed and are defined by the Department of Defense and by each of the service
branches in a series of documents and manuals listed in Appendix B. We organize and
summarize the information in these manuals in an unpublished draft document for our
project “A Summary of Military Enlistment Standards” by Christopher Buck. The standards
can be roughly categorized into 12 broad areas shown in this table. In the next several tables,
we briefly describe the standards in categories that are relevant to the study’s main findings.
12
What Factors Affect the Military Enlistment of Hispanic Youth?
Army Is the Most Lenient with Regard
to AFQT, Age, and Financial Debt
Army
Air
Force
Navy
Marine
Corps
Min. AFQT Tier 1 (HSD)
16
40
31
21
Min. AFQT Tier 2 (GED,
home school)
31
50
50
31
Min. AFQT Tier 3 (all
others)
31
65
50
50
Maximum age
34
27
28
34
Max. dependents, if
married
3
3
1
1
No
specific
policy
Not to
exceed
40% of
salary
Not to
exceed
50% of
salary
Must meet
current
financial
obligations
Financial debt
-10
This table compares the services’ eligibility requirements for education and the
AFQT score, as well as the maximum age, number of dependents, and financial debt the
military applicant can carry. Education and AFQT score are defined as tiers. Those with at
least a high school diploma (HSD) must score at least 40 on the AFQT if they are to join the
Air Force. The AFQT score is a linear combination of math and verbal test scores that are
standardized to the national 18- to 23-year-old population. Those joining the Army can have
an AFQT score as low as 16. Applicants with no high school degree, tier 3 applicants, must
have a score of 50 for the Marine Corps and Navy and a score of 31 for the Army. Enlistees
must be between 17 (with parental consent) and 35 years old to join the military according
to Title 10 of the U.S. Code. Beyond this requirement, the services may set their own
maximum. The Army and Marine Corps allow applicants as old as 34, while the Air Force
and Navy set their maximum at a lower age. The Marine Corps and Navy permit applicants
to have a spouse but no other dependents, while both the Army and Air Force allow
applicants to have a spouse and up to two children. Single parents are disqualified. Finally,
the services differ with regard to financial debt. The Air Force requires that applicants’ debt
not exceed 40 percent of their military pay. The figure for the Navy is 50 percent. Although
the Army checks applicants’ financial health, it does not have a specific policy about how
much debt load is too much.
Year One, Task One: Cataloging Current Enlistment Standards
13
Marine Corps Has the Most Lenient
Maximum Weight Standard
Maximum Weight for Males (lbs)
Navy
Marine Corps
Delayed Entry
Programa
Marine Corps
Shipb
169
165
185–190
174
189–195
194
191
215–220
202
72
200–206
205
201
227–232
213
75
217–224
224
216
246–253
231
Armya
Air
Force
65
163–168
70
Height
(inches)
NOTE:
a
Standards depend on age.
For Marine Corps, if enlistee weighs less than indicated weight and passes strength
test, then no waiver is required.
b
-11
The Army defines its weight standards in terms of height and age, while both the Air
Force and Navy base their standards on height alone. The Marine Corps also has weight
standards that vary by height and age for those entering the delayed-entry pool and signing a
contract to enter the Marine Corps at a later date. However, at entry, the Marine Corps sets
a more stringent weight standard for those who ship to recruit training. This ship weight
standard varies only by height, not by age. The Marine Corps also requires that those who
ship to recruit pass a strength requirement within 30 days prior to entry. Individuals who
exceed the Marine Corps ship weight and are below the weight required to sign a contract
and enter the Delayed Entry Program may be eligible for a waiver. A waiver requires that
they pass a strength test and possibly a body fat analysis. In the case of the Navy, those who
exceed the weight standard for their height may still be able to ship to recruit training if their
body fat is less than 22 percent.
14
What Factors Affect the Military Enlistment of Hispanic Youth?
Citizenship Is Not Required
• Must be a citizen, permanent resident, conditional
permanent resident (except AF), or noncitizen national
• Noncitizens have restricted job opportunities;
ineligible to be commissioned or warrant officers
• Army and Air Force disallow noncitizens to reenlist
• Marine Corps: Noncitizens from hostile countries
cannot have more than two visits home in five years
(except vacations and sporting events)
-12
To be eligible to enlist, an individual must be a U.S. citizen, or an alien lawfully
admitted to the United States for permanent residence, or a citizen of the Federal States of
Micronesia or the Republic of the Marshall Islands, or a noncitizen national. Included
among permanent residents are individuals admitted as conditional permanent residents, i.e.,
individuals whose residence status is subject to expiration after two years. Noncitizen
nationals are individuals born in American Samoa.
Though eligible for enlistment, non–U.S. citizens are restricted in how they can
serve. First, they are not eligible for appointment as commissioned or warrant officers or for
jobs that require a security clearance. Second, the Air Force does not allow noncitizens to
reenlist unless they have become citizens during their first term of service. Similarly, the
Army bars noncitizens with more than eight years of service from reenlisting. The Marine
Corps and Navy do not have such restrictions and allow all noncitizens to reenlist regardless
of whether or not they have become citizens while serving.
The Marine Corps also specifies that noncitizens who have residency in countries
considered hostile toward U.S. interests cannot have taken more than two trips in five years
after the 15th birthday preceding enlistment, except for school trips, family vacations, or
other similar trips.
Year One, Task One: Cataloging Current Enlistment Standards
15
Language Proficiency Is Required
• ASVAB is administered in English
• Language proficiency requirements vary across services
Army
Air Force
Navy
Marine Corps
Applicant
must be able
to understand
oath and
entrance
processing;
some
remediation
available
Applicants lacking
language
proficiency can
retake ASVAB
Applicant must
be proficient in
English prior to
entry; some
remediation
available
Applicant must
be proficient in
English to
complete recruit
training
-13
Because the Armed Services Vocational Aptitude Battery (ASVAB), the exam from
which the AFQT is derived, is administered in English, it operates as a screen for English
language proficiency. If a recruit lacks the language skills necessary to qualify based on
AFQT scores, he or she is disqualified. The only recruits who are accepted for enlistment
with disqualifying AFQT scores are those enlisted in the Army’s Individual Ready Reserve
(IRR) Direct Arabic Linguist Program, beginning in August 2003. This is one of three
programs run by the Army to provide English language training to recruits who need
remediation. The Navy also provides such training.
The Army attempts to identify all recruits who meet AFQT standards but who have
difficulties speaking or understanding English. These recruits are given the English Language
Comprehension Level Test (ELCLT) at the Military Entrance Processing Station (MEPS),
and those who score poorly are required to attend an English language training program.
The Air Force also uses the ELCLT but only at the MEPS in Puerto Rico. The Navy
considers the applicants’ verbal ASVAB scores. Those who score poorly are required to take
an additional test, the Test of Adult Basic Education (TABE). Those recruits who score
poorly are provided either literacy or English language training, depending on their needs.
The Army runs three English language programs for enlisted recruits. The first is a
traditional English-as-a-second-language program. Recruits are not assigned Army
occupations, and are permitted to retake the ASVAB following completion of the program
and be assigned their occupations once their new ASVAB scores are known. The second
program was started in 2001 and is a pilot program that uses a Spanish language screening
test to identify high-aptitude Spanish speakers who score poorly on the AFQT. Participants
are permitted to retake the ASVAB following completion of the program. The third program
is the IRR Direct Arabic Linguist Program.
16
What Factors Affect the Military Enlistment of Hispanic Youth?
Moral Character Standards
• Each applicant requires an ENTNAC
• Categories of legal infractions
•
•
1. Minor traffic
2. Minor nontraffic
3. Misdemeanors
4. Felonies
The number of infractions permitted before waiver varies by
service
Nonwaivable standards include the following:
– Convictions for drug trafficking, sales, or distribution
– Pending judicial proceedings or charges
– Civil restraints including confinement, parole, probation
– Felonies while in DEP
-14
The services also require that recruits meet moral character standards where the
standards refer to past legal infractions and substance abuse. Each service defines different
categories of infractions and sets limits on the number of offenses for each category before an
applicant requires a waiver or be disqualified for service. The categories are quite broad and
cover minor traffic offenses, nontraffic offenses, and so forth. However, the particular
offenses within each category are quite detailed and differ across service branches. For
example, the Army will permit the enlistment of an applicant convicted of a misdemeanor,
while the other service branches may require a waiver or consider such an applicant
disqualified. To ensure that all violations are included when assessing moral eligibility, an
Entrance National Agency Check (ENTNAC) is conducted for all applicants.
Some offenses are considered nonwaivable. These include drug trafficking, sales, and
distribution, and for the Marine Corps, felonies while in the Delayed Entry Program (DEP).
Individuals on parole, probation, or civil confinement are also not eligible for enlistment, nor
are those with pending judicial proceedings or awaiting criminal charges.
Year One, Task One: Cataloging Current Enlistment Standards
17
Substance Use Standards Vary by Service
Offense
Preservice use
of marijuana
Preservice use of drugs
other than marijuana
Drug or alcohol
dependency
Drug
trafficking
Must disclose
use; no waiver
required
Must disclose use; no
waiver required
Not waivable
Army
Not
waivable
Air Force
5 or more uses
requires
waiver
Use of amphetamines,
barbiturates, and anabolic
or androgenic steroids are
waivable
Waivable if abstained
from alcohol for 2
years prior to
enlistment, or in rehab
for marijuana
Not
waivable
Navy
No waiver
required for
experimental
use
Use of stimulant,
depressant, narcotic, and
psychedelic drugs is
waivable if not used in
previous year; 2 years for
LSD
Prior psychological or
physical dependence
upon any drug or
alcohol is waivable
Not
waivable
1 or more use
requires a
waiver
1 or more use requires a
waiver
Not waivable
Not
waivable
Marine
Corps
-15
In general, all services disqualify applicants who are currently dependent on drugs or
alcohol. Drug and alcohol testing are part of the physical examination applicants undergo at
the MEPS. The services differ somewhat on what past drug- and alcohol-related activities
require a waiver. The Marine Corps and Army do not generally grant waivers to applicants
who have a history of drug or alcohol dependency or addiction. The Navy and Air Force
may grant waivers in some circumstances. For marijuana use, the Army states that applicants
with a history of chronic use are disqualified, the Air Force requires a waiver for applicants
who have used marijuana more than five times before enlistment, and the Navy does not
require a waiver for casual use of marijuana. The Marine Corps has the strictest policy and
requires a waiver for all applicants who have ever used marijuana. The second phase of our
research is investigating the propensity of the services to grant enlistment waivers.
CHAPTER THREE
Year One, Task Two: Assessing the Importance of Enlistment
Standard Characteristics Among Hispanics
Research Tasks
Year 1 • Task 1: Catalog current enlistment standards
• Task 2: Assess the importance of enlistment standard
characteristics among Hispanics
• Task 3: Examine how the fraction passing enlistment
standards has changed over time
Year 2 • Task 4: Estimate the representation of Hispanics in
military accession by Hispanic subgroup
• Task 5: Analyze the career and schooling paths of
Hispanics
• Task 6: Analyze MEPCOM attrition among Hispanic
applicants
• Task 7: Analyze the waiver process and the effects of
enlistment standards on the success of Hispanic recruits
during their first enlistment term
-16
The remainder of the briefing presents our analysis of the fraction of the Hispanic
youth population who would meet these enlistment standards and how it has changed over
time.
19
20
What Factors Affect the Military Enlistment of Hispanic Youth?
Estimating the Fraction Who Meet Enlistment
Standards Is Complicated by Two Factors
1. Enlistment standards are fluid due to waivers
• Waivers can be granted at different organizational levels
2. Enlistment standards do not match directly to variables measured in our
data
• No one data set contains information relevant to the entire set of
enlistment standards
• We rely on four different data sets:
– National Health Interview Survey (NHIS) (National Center for
Health Statistics, 1998, 1999, 2000, 2001)
– 1979 National Longitudinal Survey of Youth (NLSY) (U.S.
Department of Labor, 1979)
– 1997 NLSY (U.S. Department of Labor, 1997)
– June 2003 Youth Poll (YP) Survey (Defense Human Resources
Activity, 2003)
-17
We faced two challenges in estimating the fraction of the Hispanic youth population
who would meet enlistment standards. First, enlistment standards are not always hard and
fast because applicants who do not meet a specific standard can sometimes receive a waiver.
For example, a highly qualified Navy applicant with more than six minor traffic violations
could receive a waiver, if the violations occurred in the distant past or if the Navy was in
great need of recruits to meet its monthly recruiting mission. The waiver process is not
entirely transparent; it depends on the decision of the commander at the local level or higher,
depending on the nature of the waiver (e.g., a minor traffic violation or a conviction). It also
seems to depend on the recruiting needs of the service. Tasks 6 and 7 of the project in the
second year will be investigating waivers in greater depth. The implications of waivers is that
we will tend to overestimate the fraction of youth who will not meet any given standard, as
some youth who appear unqualified might receive waivers.
The second challenge is that the standards are quite detailed requirements, while the
variables in the data used in our analysis are defined broadly and do not perfectly measure
the criteria applied by each service. Furthermore, no single data source contains information
on all enlistment standards. We therefore rely on multiple data sources, and the measures
used in them are those that, on balance, most closely capture the set of qualifying standards.
Thus, the measures provide an estimate of the fraction of the population that is likely to
qualify for enlistment. Also, we compare the fraction likely to qualify among the Hispanic
population relative to other demographic groups. That is, we focus more on comparisons of
the fraction who qualify across groups rather than on the absolute magnitude of the estimate
of the fraction who qualify.
The data sources we use to examine medical standards are the 1998–2001 waves of
the National Health Interview Survey (NHIS) (National Center for Health Statistics, 1998,
1999, 2000, 2001). For most of the remainder of the standards, we use the 1997 National
Year One, Task Two: Assessing the Importance of Enlistment Standard Characteristics Among Hispanics
21
Longitudinal Survey of Youth (NLSY) (U.S. Department of Labor, 1997). Unlike any other
data source, the NLSY has recent information on AFQT scores. It also has information on
education, number of dependents, height and weight, and some information on criminal
activity. To examine how the fraction who qualify has changed over time, we compare results
between the 1979 and 1997 NLSY (U.S. Department of Labor, 1979, 1997). Finally, in
June 2003, the Department of Defense Joint Advertising Market Research and Studies
Program conducted a survey of youth to examine enlistment eligibility explicitly (Defense
Human Resources Activity, 2003). Although it does not have information on AFQT, this
survey provided more detailed information on various moral character-related standards. A
detailed discussion of the translation of the enlistment eligibility criteria into variable
measures from each data source is discussed in subsequent charts.
Finally, documentation from all four data sources indicates that information on
whether a respondent is of Hispanic origin is based on self-identification, either from the
respondent, or in the case of minors in the NLSY, from a parent or guardian. That is, the
information is based on the individual’s response to a question about his or her ethnicity or
about whether he or she is Hispanic. Information on race is also based on self-identification,
with the exception of the 1979 NLSY. In that survey, race is based on the interviewer’s
observation. It should be noted that there are no missing values in the NLSY or NHIS, and
the interviewer or organizations that conducted the surveys provide values in the missing
cases, by observation, or through imputation. Unfortunately, because of a lack of sample size,
we could not perform our analysis by Hispanic subgroup. The exception is the NHIS
analysis, which considers all Hispanics, then the subgroup of those of Mexican origin only.
22
What Factors Affect the Military Enlistment of Hispanic Youth?
We Use the National Health Interview Survey
(NHIS) to Study Health-Related
Disqualifications
• Pool 1998–2001 survey waves
– Adults age 18–30: 12,013 males, 14,792 females
• Examine three types of health disqualifications
based on likelihood of waiver
– Weight-for-height
– Major health conditions (e.g., functional
limitation, heart condition, blind or deaf, special
equipment, organ failure, heart conditions,
ulcer, stroke, active asthma)
– “Minor” health conditions (e.g., chronic
bronchitis, hay fever, sinusitis, cancer)
-18
The NHIS is an annual cross-sectional, household-based survey of diseases, health
conditions, and behaviors in the civilian population. One adult (age 18 and over) is sampled
from each family and asked detailed demographic and health questions. The survey
oversampled blacks and Hispanics. By pooling several years together, 1998–2001, we are able
to analyze the Mexican sample as a distinct subsample from the general Hispanic sample
with statistical precision. From the adult sample, we consider males and females 18–30.
Because children under 17 are sampled separately and asked different health questions, we
limit the current briefing to the results for adults. We use a wider age range than the
military’s prime recruiting target range of 17 to 24 to increase sample size. The older ages are
relevant because both the Army and Marine Corps permit recruits as old as 35.
The goal of the analysis is to estimate the fraction of the youth population who
would meet the military’s health and medical standards, by race and ethnicity. The analysis is
described in detail in an unpublished draft document for our project “Health Barriers to the
Enlistment of Hispanic Youth,” by Meredith Kleykamp. The analysis is descriptive and is
driven by the NHIS data definitions. Measuring the prevalence of enlistment-related health
disqualifications is challenging for reasons given earlier. A military applicant is screened by a
doctor who not only has the patient with whom to interact and physically examine, but who
also has past medical records at his or her disposal. As a result, the services can delineate
specific health conditions. Large national data sets employ coarser measures.
The strategy adopted in this study is to identify three general health measures,
indicating whether an individual would meet the health standards on the basis of weight or
of a major or minor health condition. This measurement strategy captures most of those who
would fail to meet the medical enlistment standards across the services. Further, the strategy
of aggregating individual conditions is used because of the relative infrequency of any single
disqualifying health condition. There are simply not enough cases to examine each condition
Year One, Task Two: Assessing the Importance of Enlistment Standard Characteristics Among Hispanics
23
separately with any precision. In the case of weight, self-reported height and weight are
compared to the sex- and age-specific height-weight accession standard for each service.
We define major conditions as those that the military treats as nonwaivable
(according to the published medical standards for each of the four branches). Major
conditions are defined specifically as follows: having a functional limitation; having a
noncorrectable vision or hearing problem; needing special equipment for daily activity;
having organ failure, heart condition, diabetes, ulcer, stroke, emphysema, or hypertension; or
having an asthma attack in the past year. This last measure of asthma does not perfectly
capture the military standard for asthma. For the military, active asthma after age 14 is
considered nonwaivable. But asthma before age 14 may be waivable (i.e., it may require a
pulmonary function test or other tests or medical records to waiver). Our measures only ask,
“Have you ever been told by a doctor you have asthma?” or “During the past 12 months,
have you had an episode of asthma or asthma attack?” We use information about recent
attacks, treating this as an indicator of certain disqualification from enlistment, rather than
using information about having ever been diagnosed with asthma, as many individuals may
have been diagnosed with asthma as a child, but have not had active asthma since childhood.
This potentially underestimates the number of individuals who might be disqualified from
asthma.
Minor conditions are defined as disqualifying, but waivable. These conditions
include sinusitis, hay fever, and chronic bronchitis. Also included are conditions where our
metric does not precisely identify waivability. Cancer is the only condition in this category.
Our measure asks whether an individual has ever had cancer. Having ever had cancer does
not automatically disqualify a person from enlistment. Individuals whose cancer has been in
remission for five years or longer are eligible for enlistment. Unfortunately, we cannot
identify if a cancer is in remission or the duration of remission. Again, because of the
imprecise matching of the condition with the measure, we treat it as potentially waivable.
Because the incidence of cancer is so low in the population being studied, whether included
as a minor or major condition, it would not have a large influence on the total.
Several analyses are performed. First, we estimate the fraction of non-Hispanic
whites, non-Hispanic blacks, and Mexicans who would meet the health standards.1 In
addition, the results for an aggregate Hispanic category (that includes Mexicans) are also
presented. Those who join the military are not a random sample of the population. To
account for this, the simple cross-tabulation strategy is performed on successively restrictive
populations, focusing on the “military preferred” population. The military preferred
population refers to those who are U.S. citizens and high school graduates.
The health condition questions in the NHIS are mostly self-reports of having ever
been diagnosed with a given condition by a doctor. Consequently, this measure may not
accurately capture the prevalence of a condition if there is underdiagnosis in certain
populations for any reason. This is especially worrisome for the Hispanic population, which
has much lower rates of insurance, and may face language and other barriers to getting
medical care. As a consequence, individuals may have a condition that is undiagnosed either
____________
1
Other Hispanic ethnic subgroups, such as Cubans or Puerto Ricans, simply do not have enough cases, even pooling
multiple years of data, to give statistically precise estimates, given our strategy of successively restricting the sample. The
aggregate Hispanic category includes Mexicans, Puerto Ricans, Cubans, and Other Hispanics. Therefore, the Hispanic
category and Mexican category are not mutually exclusive. As a consequence, estimates for Mexicans and Hispanics come
from two separate models.
24
What Factors Affect the Military Enlistment of Hispanic Youth?
because they have not seen a doctor recently or because they have not had high-quality care.
To try and address this concern, the results presented for health conditions are adjusted for
having seen a doctor in the past 12 months. The adjustment strategy is a simple regressionbased adjustment. First, the regression is estimated including ethnicity dummies and a
doctor visit indicator. Then, the model coefficients are used to generate predicted
probabilities, treating all individuals as having seen a physician in the past year, and then
summarizing these predicted probabilities by ethnicity. This strategy is used for the estimates
of both major and minor conditions for the adult samples. We also analyzed the results (not
shown here) without adjusting for having seen a doctor in the past 12 months. We found the
prevalence of disqualifying conditions were lower by about two to five percentage points, for
all ethnic and racial groups. The ordering of the prevalence of disqualifying conditions across
groups did not change when no adjustment was made. Thus, the adjustment affects the
absolute magnitude of the estimates for each group but not the comparisons across ethnic
and racial groups.
Year One, Task Two: Assessing the Importance of Enlistment Standard Characteristics Among Hispanics
25
Findings on the Health of Hispanic Youth
from Previous Research Are Mixed
• On average, Hispanics are poorer and less educated than
whites; less educated than blacks
– Income and education are positively related to health
status
• In terms of specific health conditions, Hispanics are:
– More likely to have infectious diseases, diabetes, obesity
– Less likely to have cancer, low birth weight
• “Immigrant paradox” is found in Hispanic population
– Hispanics have better-than-expected health outcomes,
especially infant mortality and birth outcomes, and adult
mortality rates
– Cause of paradox debated, but immigrants appear to have
better health outcomes than native-born Hispanics, except
for diabetes and weight
-19
Previous research on the health of Hispanics has found mixed results. The findings
depend largely on the population chosen for analysis or the specific health condition
measured. On the one hand, there is reason to suspect Hispanics may suffer a greater burden
from many of the health conditions considered for enlistment. Hispanics are on average
poorer and less educated than whites, and less educated on average than blacks (Hoffman,
Llagas, and Snyder, 2003; Llagas and Snyder, 2003). Because income and education are
positively correlated with health, we would expect Hispanic health patterns to reflect their
relative socioeconomic patterns. Infectious diseases, diabetes, and obesity disproportionately
affect Hispanic groups, but for cancer, low birth weight, and some other health outcomes,
Hispanics do not appear to have a disadvantage (see Vega and Amaro [1994] for a thorough
review). Numerous studies have documented higher levels of overweight status and obesity
among Hispanics, even among military personnel (National Center for Health Statistics,
2003; Lindquist and Bray, 2001; Nolte et al., 2002; Popkin and Udry, 1998).
On the other hand, researchers have identified a paradox of better-than-expected
health outcomes (especially for mortality, infant mortality, and birth outcomes) in the adult
Hispanic population, often referred to as the “Hispanic paradox” or “epidemiologic
paradox.” The true cause of the paradox is debated, but research has generally concluded that
the large fraction of immigrants in the Hispanic population is the primary explanation for
the Hispanic paradox (now often referred to as the “immigrant paradox”).2 In fact, recent
____________
2 The
two main explanations for the paradox center on the selectivity of migrantion and on the effects of acculturation. It is
hypothesized that only the healthiest individuals can undertake moving to another country, legally or especially illegally,
resulting in immigrants being positively selected on health. The acculturation hypothesis suggests that immigrants have
healthier lifestyles, eschewing smoking, alcohol, and unhealthy “Americanized” diets. The longer they live in the United
States, the more acculturated they become, taking on more unhealthy behaviors. There is evidence to support both
26
What Factors Affect the Military Enlistment of Hispanic Youth?
studies have found a paradox for most ethnic groups when comparing immigrant to
nonimmigrant health outcomes with weight and diabetes being the notable exceptions to the
health paradox for Hispanics (Cho and Hummer, 2001; Franzini, Ribble, and Keddie, 2001;
Frisbie, Cho, and Hummer, 2001; Palloni and Morenoff, 2001). Recent research has also
identified a similar pattern of better-than-expected health among first-generation Mexican
children in the United States (Burgos et al., 2005; Mendoza and Fuentes-Afflick, 1999).
Given these alternative expectations about the relative health status of Hispanics compared
with blacks or whites, it is an empirical question whether health may serve as an obstacle to
enlistment for Hispanics.
________________________________________________________________________
hypotheses, and is often highly dependent upon the ages considered and the specific health measure being used. (See in
particular Franzini, Ribble, and Keddie [2001] and Palloni and Morenoff [2001]).
Year One, Task Two: Assessing the Importance of Enlistment Standard Characteristics Among Hispanics
27
Mexican Males Are the Least Likely
Group to Meet Weight Standards
Percent of males meeting weight standards of at least
one branch
100%
90%
Age 18–30
80%
70%
Age 18–25
60%
50%
40%
U.S. citizen
30%
20%
H.S. graduate
10%
0%
White
Black
Mexican
All Hispanic
-20
The first bar for each racial or ethnicity group represents the full sample of 18- to 30year-old males for each group. Next, the population is limited to a slightly younger group,
18- to 25-year-old males. The third bar represents only those 18- to 25-year-old males who
are U.S. citizens. Next, the population is further limited to 18- to 25-year-old male citizens
who are high school graduates. This same logic applies to all of the charts presented. As
noted earlier, the 18–25 group and citizens are the “military preferred” population.
There are two types of information presented in these charts—information on ethnic
differences within subpopulations and differences from the selectivity of subpopulations
within ethnic group. One can identify ethnic differentials by comparing the same shaded
bars across race and ethnicity groups. Comparing across bars within each group reveals how
subpopulation selection affects the prevalence of various health conditions.
We find that Mexicans are the most likely to have difficulty meeting the weight
requirements, with the pattern for all Hispanics roughly mirroring the results for Mexicans.
In any given population, Mexicans report the lowest fraction of males who would meet the
weight standards of one or more branches of the military, with about 60 to 70 percent of
Mexican males meeting the weight standards or as many as 30 to 40 percent of Mexican
youth failing the weight standards for enlistment in at least one branch—a fairly staggering
prospect. As we restrict the population to a younger subset, ages 18 to 25, the fraction who
meet the weight standard rises for all groups, but falls, except for whites, as we further restrict
to the military preferred population of U.S. citizens and high school graduates, indicating
that those the military prefers are less likely to pass the weight standards. Selection appears to
act most substantially at the citizenship criteria for Mexicans. Thus, Hispanics who are
immigrants are more likely to pass the weight standard. Selection occurs more at the high
school graduate level for blacks.
28
What Factors Affect the Military Enlistment of Hispanic Youth?
Black Females Are the Least Likely Group
to Meet Weight Standards—Adjusted
Percent of females meeting weight standards of at
least one branch
100%
90%
Age 18–30
80%
70%
60%
Age 18–25
50%
40%
U.S. citizen
30%
20%
H.S. graduate
10%
0%
White
Black
Mexican
All Hispanic
-21
Unlike males, we find that black females are the most likely to face weight barriers to
enlistment. Fewer than 40 percent of black young women would meet the weight standards
of at least one branch. White females are the least likely to face difficulties meeting the
weight standard, with roughly half of all white females meeting the standards set by at least
one branch. Mexicans and all Hispanics lie somewhere in between, with approximately
41–45 percent meeting the standards. Further, it does not appear that citizenship has as
strong an effect as for Hispanic males. In all groups, high school graduates appear to be more
likely to meet the weight standards, a result unlike that for males. The differences across
racial and ethnic categories and population segments pale in comparison to the striking result
that only about a half or fewer women would meet the weight standards of at least one
branch of the military; clearly weight is a larger obstacle to military service for females than
for males.
Year One, Task Two: Assessing the Importance of Enlistment Standard Characteristics Among Hispanics
29
A Substantial Fraction of Overweight
Hispanic Males Is Within a Few Pounds
of the Weight Standard
White
Black
Mexican
Hispanic
5 lbs. overweight
16%
15%
16%
19%
10 lbs. overweight
30%
26%
36%
37%
-22
Given the large fraction of men, both Hispanic and non-Hispanic, who would fail to
meet the military’s weight standard, additional analyses were conducted to evaluate the
distribution of individuals around the weight cutoff points. That is, we asked whether a large
or small number of individuals are within a few pounds of the weight standards, and whether
this differed by race or ethnicity. If so, a slight relaxation of weight standards could result in
more Hispanics being eligible for service. Kernel density plots were made of the difference
between observed weight and the maximum allowable weight, by race or ethnicity. These
density plots are summarized by examining the fraction of individuals failing the military
weight standards who are over the maximum weight by five pounds or less and by ten
pounds or less. This chart summarizes the results.
About 16 percent of males 18 to 25 are overweight by five pounds or less
(conditional on being over the maximum standard weight). The percentage is slightly higher
for Hispanic males (19 percent) and lower for black males (15 percent). Roughly a third, 37
percent, of Hispanic males are within 10 pounds or less of the standard (conditional on
being over the maximum standard weight). The figures are lower for white and black males,
30 percent and 26 percent, respectively. These figures suggest that a nontrivial fraction of
men would qualify for service on the basis of their weights, if the weight standard were
relaxed by a few pounds.
30
What Factors Affect the Military Enlistment of Hispanic Youth?
A Substantial Fraction of Overweight
Hispanic Females Is Within a Few
Pounds of the Weight Standard
White
Black
Mexican
Hispanic
5 lbs. overweight
15%
11%
14%
14%
10 lbs. overweight
27%
23%
27%
27%
-23
Slightly fewer women than men fall within five or ten pounds of the weight
standards. About 14 percent of females 18 to 25 are overweight by five pounds or less
(conditional on being over the maximum standard weight). The percentage is slightly higher
for white females (15 percent) and lower for black females (11 percent). Approximately a
quarter, 27 percent, of Hispanic females are within 10 pounds or less of the standard
(conditional on being over the maximum standard weight). The figures are lower for black
females, with 23 percent falling within 10 pounds of the limit. These figures suggest that,
though fewer than men, a nontrivial fraction of women would also qualify for service on the
basis of their weight, if the weight standard were relaxed by a few pounds.
Year One, Task Two: Assessing the Importance of Enlistment Standard Characteristics Among Hispanics
31
Mexican Males Are Healthier in Terms
of Major and Minor Qualifying
Conditions
Percent of males without a major or minor disqualifying
health condition
100%
90%
80%
Age 18–30
70%
60%
Age 18–25
50%
40%
U.S. citizen
30%
20%
H.S. graduate
10%
0%
White
Black
Mexican
All Hispanic
-24
When we aggregate, asking whether an individual has a major or minor condition,
we find that Mexicans are less likely to have a health-disqualifying condition. That is, they
are more likely to meet the health standards of at least one service for these conditions. The
pattern of selectivity as we define the sample in terms of military preferred (18 to 25, U.S.
citizen, high school graduate) shows a large drop in the likelihood of meeting health
standards among U.S. citizens for Mexicans and Hispanics. That is, citizens are not as
healthy as noncitizens. For blacks, high school graduates appear to be slightly less healthy
than the average young man.
32
What Factors Affect the Military Enlistment of Hispanic Youth?
Mexican Females Are Healthier in
Terms of Major and Minor Qualifying
Conditions
Percent of females without a major or minor disqualifying
health condition
100%
90%
Age 18–30
80%
70%
Age 18–25
60%
50%
U.S. citizen
40%
30%
H.S.
graduate
20%
10%
0%
White
Black
Mexican
All Hispanic
-25
When looking at the percent of females who have a major or minor health condition,
we find that Mexican females are also less likely to have a health disqualifying condition.
That is, they are more likely to meet the health standards of at least one service for these
conditions. The pattern of selectivity as we define the sample in terms of military preferred
(18 to 25, U.S. citizen, high school graduate) also shows a drop in the likelihood of meeting
health standards among U.S. citizens for Mexicans and Hispanics. That is, citizens are not as
healthy as noncitizens. Unlike the results for males, it appears that white females are the least
likely overall to meet the standards for this set of health conditions.
Year One, Task Two: Assessing the Importance of Enlistment Standard Characteristics Among Hispanics
33
Overall, Hispanic Males Are as Likely as
Whites and More Likely than Blacks to
Meet Health Standards
Percent of males without any disqualifying health
condition (major, minor, or weight disqualification)
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
Age 18–30
Age 18–25
U.S. citizen
H.S. graduate
White
Black
Mexican
All
Hispanic
-26
This chart shows the fraction of each group without any of the disqualifying
characteristics (weight, major, or minor) for at least one service. Because of the higher
prevalence of weight problems among young men of Hispanic or Mexican origin, but lower
prevalence of other major or minor health conditions, we see how there is little net difference
between whites and Hispanics when aggregating these measures in the fraction of Hispanic
young men who meet the health standards. This result holds for the “military preferred”
population, but among noncitizens, Mexican and Hispanic males have lower overall rates of
health disqualification. The effect of considering all health conditions, including weight, is
striking. The only real difference for Hispanics from the earlier chart is in the fraction who
meet the standard—which is reduced by roughly 20 to 25 percentage points when we
consider all health conditions together.
34
What Factors Affect the Military Enlistment of Hispanic Youth?
Overall, Hispanic Females Are as Likely
as Whites and More Likely than Blacks
to Meet Health Standards
Percent of females without any disqualifying health
condition (major, minor, or weight disqualification)
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
Age 18–30
Age 18–25
U.S. citizen
H.S. graduate
White
Black
Mexican
All
Hispanic
-27
When examining females, we see Hispanic females are slightly more likely to meet
the total set of health standards set by the military. Because of the high prevalence of weight
problems among black women, and their poorer health status compared with Hispanics,
black women are the least likely to meet the set of health standards considered. The effect of
considering all health conditions for any of the groups, including weight, is striking. The
only real difference for Hispanics from the earlier chart is in the fraction of women who meet
the standard—which is reduced by roughly 30 to 40 percentage points when we consider all
health conditions together.
Year One, Task Two: Assessing the Importance of Enlistment Standard Characteristics Among Hispanics
35
Summary of Findings on Health-Related
Enlistment Standards
• Hispanic and black males and females are the least likely to
meet weight standards
• Hispanic males and females are more likely to meet other
health-related standards
• Net effect: Hispanics are as likely as whites and more likely
than blacks to meet all three health-related categories we
considered
• Across all groups, females are less likely to meet health
standards
– Weight presents a much greater obstacle for females than
males for all racial and ethnic groups
-28
To summarize our conclusions from our analysis of health conditions from the
NHIS, we find that Hispanic and Mexican males are less likely to meet the weight standards
than their black or white counterparts, but that black females are the least likely to meet the
weight standards, followed by Mexican and Hispanic females. However, both male and
female Mexicans and Hispanics appear to be healthier when considering other medical
qualification standards—having lower prevalence of both the major disqualifying conditions,
and of the minor or potentially waivable conditions. When aggregating, these divergent
results cancel out, and seem to indicate that, overall, health is not the likely cause of
underrepresentation. It is also clear that there is a strong selection effect. In general, citizens
of Mexican origin and black high school graduates are less likely to meet the health
standards. The large effect of citizenship on the prevalence of disqualifying health conditions
for Mexicans is consistent with other analyses showing immigrants are healthier than nativeborn peers.
36
What Factors Affect the Military Enlistment of Hispanic Youth?
We Use the 1979 and 1997 NLSY Cohorts
and the 2003 Youth Poll to Examine
Additional Disqualifications
• 1979 and 1997 National Longitudinal Survey of
Youth (NLSY79 and NLSY97)
– Ages 17–21 in 1983 or 2001
– Estimate fraction passing standards related to
high school graduation, AFQT, height and
weight, dependents, and criminal activity
• 2003 Youth Poll
– Ages 17–21
– Youth Poll has better questions related to
criminal activity
-29
We next turn to our analysis of non–health-related enlistment standards using the
1997 NLSY and 2003 Youth Poll. The NLSY is a nationally representative sample of
American youth ages 12 to 16 as of December 1996 who are interviewed on an annual basis.
We limit our analysis to the 2001 responses, when the cohort was ages 17 to 21 and
therefore in the target recruiting age range. The sample includes both males and females, for
a sample size of 6,180 observations. We estimate the fraction of the youth population who
would pass each service’s standards, by race and ethnicity. We also use the 1979 NLSY to
examine how the fraction of youth passing enlistment standards has changed over time.
The NLSY’s variable definitions drive the definition of the standards in our analysis.
In some cases, such as education and AFQT score, we can often directly match the variable
measure with the definition of the standard. In other cases, notably in the case of the moral
character standards relating to drug use and convictions, the variable measures are far broader
than the specific enlistment standards defined by the services. We define the standards as
follows with the NLSY data.
For the purposes of this study, we define a high school graduate as someone who has
either a traditional high school diploma or a General Equivalency Diploma (GED). DoD
generally prefers traditional high school graduates to those with a GED, but can legally
accept those with a GED providing they score sufficiently highly on the AFQT. Since the
GED population is so small, restricting our definition of a high school graduate to just those
individuals who have a traditional high school diploma has virtually no effect on the results
we report below. AFQT score is defined as meeting the absolute minimum score. For
example, in the case of the Army, the score is 16. In the case of the Air Force, the score is 40.
In charts that show the cumulative effect of enlistment standards, AFQT is defined
conditionally on the type of diploma a given individual holds. “Dependents” refers to
meeting the service’s standard for numbers of dependents. “Weight” is defined as meeting
Year One, Task Two: Assessing the Importance of Enlistment Standard Characteristics Among Hispanics
37
the service’s shipping weight requirements. “Convictions” is defined as ever having been
convicted of a crime. “Drug use” is defined as having used marijuana or other drug use in the
last year.
The Youth Poll is a computer-assisted telephone interview survey of youth ages 16 to
21, conducted in spring 2003. We limited the analysis to respondents ages 17 to 21. A big
advantage of the Youth Poll is that its survey questions are posed in a way that is directly tied
to how enlistment standards are measured. Thus, the survey asks, “Have you ever been
convicted of a misdemeanor? If yes, how many?” It also queries respondents about height
and weight, but does not have information about AFQT, a key variable in our analysis.
38
What Factors Affect the Military Enlistment of Hispanic Youth?
Hispanic High School Graduation Rates Differ
Between the NLSY97 and Census
•
•
•
•
•
Fraction of Hispanics 18–21, not currently enrolled in high school with high
school diploma or GED:
– 2001 NLSY97: 79 percent
– 2000 Census: 59 percent
Differences in high school graduation rates between census and NLSY97 are
not observed for non-Hispanic whites and blacks
Differences in immigration patterns cannot explain these high school
graduation rate differences
A possible explanation
– The NLSY97 is a more demanding and invasive survey than the census
– The types of recent immigrants willing to participate in the NLSY97 might
be very different from the types of recent immigrants willing to participate
in the census
– These differences in types could be correlated with educational
attainment
More research is needed to understand why Hispanic high school graduation
rates differ so sharply between the census and NLSY97
-30
As stated at the outset, failure to graduate from high school is an important reason
that many Hispanic youth do not qualify for military service relative to non-Hispanic
groups. An interesting anomaly, however, is that the high school graduation rates in the
NLSY97 are substantially higher relative to the rates in the census and the Current
Population Survey, the two standard (and considered to be highly reliable) sources of
information on the U.S. population. The fraction of Hispanics ages 18 to 21 not currently
attending high school who have a high school diploma or GED in 2001 is 59 percent in the
2000 Census but 79 percent in the NLSY97. High school graduation rates of non-Hispanic
whites and blacks are the same in the two surveys.
A much higher fraction of Hispanics in the census are recent immigrants, which
might explain differences in high school graduation rates. But, even when we look at just
recent Hispanic immigrants in both surveys, differences in high school graduation rates
remain. We conjecture, then, that recent Hispanic immigrants in the census and NLSY97
differ along dimensions that we cannot observe in one or both surveys. Differences in
country of origin, which are not well measured in the NLSY97, is one possibility. It is also
possible that recent Hispanic immigrants willing to participate in the NLSY97 on average
come from better-educated and higher-income families than do other recent Hispanic
immigrants. Willingness to participate in the NLSY97 may be correlated with certain
personality characteristics and contextual factors that also correlate with educational
attainment, like willingness to trust survey personnel or living in a safe neighborhood.
Whatever the reason, it is clear that NLSY97 Hispanics are different from census
Hispanics with respect to educational attainment and it is important to keep in mind that
these differences could be correlated with other enlistment standards. We also note that these
same differences exist between Hispanics interviewed in the 2003 Youth Poll and census.
Year One, Task Two: Assessing the Importance of Enlistment Standard Characteristics Among Hispanics
39
Blacks and Hispanics Score Much Lower on
the AFQT Than Do Whites
High School Graduates
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
Cumulative fraction
Cumulative fraction
All
0
20
Hispanic
40
60
AFQT percentile
White
80
Black
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0
20
40
60
80
AFQT percentile
Hispanic
White
Black
-31
The AFQT was recently renormed by the Department of Defense in July 2004 using
the 1997 NLSY data. However, these new norms have not yet been applied to the raw test
scores available in the NLSY data. Therefore, we devised a method for norming the raw
scores that corresponds quite closely to the DoD approach, summarized in Segall (2003).
Specifically, we performed a principal components analysis on the four components of the
ASVAB that DoD uses to derive the AFQT. Principal components transform a number of
typically correlated variables into a smaller number of uncorrelated variables, called principal
components, while maintaining the variability in the original variables. Because age in the
sample at the time of the administration of the ASVAB ranged from 14 to 17, we first
adjusted each component of the ASVAB scores for age by regressing each component score
on an age in months and on an age in months squared variable. Principal components were
performed on age-adjusted ASVAB scores. We weight the principal component using the
NLSY sampling weights to arrive at a nationally normed AFQT index for the 1997 cohort.
This chart shows the cumulative distribution of AFQT scores by race or ethnicity
and high school graduation status. The left graph shows that 40 percent of white youth but
70 percent of Hispanic youth and 74 percent of black youth scores are at or below the 50th
percentile. The right graph shows that among high school graduates, the difference between
white youth and minorities narrows, but still remains quite large with 34 percent of white
youth but 62 percent of Hispanic youth and 66 percent of black youth scoring at or below
the 50th percentile.
40
What Factors Affect the Military Enlistment of Hispanic Youth?
Hispanics and Blacks Are More Concentrated
in Lower AFQT Categories
Percent Distribution of High School
Graduates
Fraction of H.S.
graduates
0.4
0.3
0.2
0.1
0
V
IVC
IVB
IVA
IIIB
IIIA
II
I
AFQT category
Hispanic
White
Black
-32
DoD categorizes potential recruits according to eight AFQT categories—V, IVC,
IVB, IVA, IIIB, IIIA, II, and I corresponding to the 9th, 15th, 20th, 30th, 49th, 64th, 92nd
and 99th percentiles of the nationally normed AFQT distribution, respectively. This chart
shows the distribution of U.S. high school graduates, ages 17 to 21, in the NLSY across the
AFQT categories, by race and ethnicity. White graduates are far more likely to be categorized
in AFQT categories I or II than either Hispanic or black graduates. Relative to black
graduates, Hispanics are somewhat less likely to be classified as a Category I, II, IIIA, or IVA,
and more likely to classified as a Category IIIB or Category V.
Year One, Task Two: Assessing the Importance of Enlistment Standard Characteristics Among Hispanics
41
For Males, Education, AFQT, and Weight
Are Obstacles to Hispanic Enlistment
Air Force
Army
1
Diploma/GED
AFQT
Ship Weight
Dependents
Conviction
Drug Use
0.8
0.6
0.4
0.2
0
White
Black
Hispanic
Fraction passing
Fraction passing
Fraction of Males Passing Enlistment Standards
1
0.8
0.6
0.4
0.2
0
White
1
0.8
0.6
0.4
0.2
0
White
Black
Hispanic
Black
Hispanic
Navy
Fraction passing
Fraction passing
Marine Corps
1
0.8
0.6
0.4
0.2
0
White
Black
Hispanic
-33
This chart shows the fraction of males in the NLSY passing the standards for each
service by race and ethnicity, to the extent that we can measure the standards. We find that
education, AFQT, and weight (the first three bars in each chart) are obstacles to enlistment
for Hispanic males, with AFQT being particularly important for both Hispanic and black
males. We find 47 percent of the Hispanic males in the NLSY sample would meet the Navy
AFQT requirement and only about 40 percent would meet the Air Force standard. The
figures for black males are 46 and 35 percent, respectively, for these services. In contrast, the
Army and Marine Corps have less stringent AFQT standards (see earlier chart) and therefore
a higher proportion of males meet their standards. For example, 71 percent of Hispanic
males meet the Army’s AFQT standard. Even so, a significant fraction of Hispanics, relative
to white males, fail to meet Army and Marine Corps AFQT standards. Also relative to white
males, of whom 85 percent are high school graduates, only 74 percent of Hispanic males and
70 percent of black males meet the services’ standards for high school graduation.
Fewer Hispanic and black males pass the weight standard than white males, although
the fraction is close in the case of the Marine Corps. For example, in the case of the Army,
70 percent of Hispanic males and 70 percent of black males would pass the standard,
compared with 77 percent of white males. In the case of the Marine Corps’ shipping
requirement, the figures are 79 percent, 77 percent, and 83 percent, respectively. The Marine
Corps has less stringent weight requirements, so a higher fraction is estimated to pass the
standard. However, we do not have information on whether individuals meet the Marine
Corps’ strength requirement. These estimates of meeting the weight standards are somewhat
greater than those found in the NHIS discussed earlier. Additional tabulations show that the
difference is attributable to age differences in the two samples. The NLSY sample includes
ages 18 to 21 while the NHIS sample includes ages 18 to 25. When we restrict the NHIS
42
What Factors Affect the Military Enlistment of Hispanic Youth?
sample to ages 18 to 21, we find that mean weight, height, and fraction with a secondary
school diploma or certificate are the same as in the NLSY sample. For example, the mean
weight and mean height are 159 pounds and 68 inches, respectively, in both samples. The
fraction with a secondary diploma is 78 percent in the NHIS sample and 78 percent in the
NLSY sample.
A significant fraction of youth report having used any drugs (both waivable and
unwaivable) in the past year, a fraction that does not vary appreciably across groups. About
34, 29, and 27 percent of white, black, and Hispanic males, respectively, report having used
drugs in the past year. This particular statistic, however, does not correspond particularly
well to the services’ enlistment standards with respect to drug use, and so it is unclear
whether this many youth would be disqualified from enlistment because of drug use.
Moreover, self-reported drug use is likely to be underreported in survey data.
Our measure of criminal activity in the NLSY—ever convicted of a felony—is also a
poor indicator of whether a given individual would pass enlistment standards related to past
criminal activity. It is worth noting, though, that the fraction of youth reporting ever having
been convicted of a felony does not vary across groups.
In a later chart, we report separate statistics from the 2003 Youth Poll related to drug
use and criminal activity that correspond more closely to enlistment standards.
Year One, Task Two: Assessing the Importance of Enlistment Standard Characteristics Among Hispanics
43
AFQT May Be a More Important Obstacle to
Enlistment Than H.S. Diploma Status in AF, Navy
Males, age 17–21
Army
Fraction passing
0.8
0.6
0.4
0.2
0
White
Fraction passing
HSD
Black
1
0.8
0.6
0.4
0.2
0
HSD
0.8
0.6
0.4
0.2
0
White
HSD + AFQT
Marine Corps
White
1
Hispanic
Black
Hispanic
HSD + AFQT
Black
HSD
Fraction passing
Fraction passing
Air Force
1
Hispanic
HSD +AFQT
Navy
1
0.8
0.6
0.4
0.2
0
White
HSD
Black
Hispanic
HSD + AFQT
-34
The previous chart showed the importance of AFQT but not its independent effect
from high school diploma status. This chart shows the fraction who would pass the
minimum AFQT standard of each service, conditional on being a high school diploma
holder, where the AFQT standard depends on which type of diploma (traditional or GED)
an individual holds. We find that AFQT has a large independent effect for Hispanic males
for the Navy and Air Force, and sizable effects for the Army and Marine Corps. For example,
the fraction of high school diploma-holding Hispanic males who would pass the Air Force’s
minimum AFQT standards for a diploma holder is only 39 percent. For black diploma
holders, the figure is 33 percent. In contrast, the percent of diploma-holding Hispanic males
who would pass the Marine Corps’ AFQT minimum is 57 percent. Thus, AFQT is an
important obstacle to enlistment, even beyond high school, for the Navy and Air Force.
44
What Factors Affect the Military Enlistment of Hispanic Youth?
Weight and Dependents Are Significant
Obstacles for Hispanic Females
Fraction of Females Passing Enlistment Standards
Army
Diploma/GED
AFQT
Weight
Dependents
Conviction
Drug Use
0.8
0.6
0.4
0.2
0
White
Black
Fraction passing
Fraction passing
Air Force
1
Hispanic
1
0.8
0.6
0.4
0.2
0
White
Black
Hispanic
Navy
1
1
Fraction passing
Fraction passing
Marine Corps
0.8
0.6
0.4
0.2
0
White
Black
Hispanic
0.8
0.6
0.4
0.2
0
White
Black
Hispanic
-35
Weight is a significant obstacle to enlistment for females. Only 42 percent of black
females and 48 percent of Hispanic females in the NLSY sample would meet the Army’s
weight standard, given their heights. Weight is also a significant obstacle for white females,
with only 62 percent meeting the Army’s standard. Unlike males, dependent status is an
important disqualification characteristic for females. As shown in the earlier chart, 90 percent
of Hispanic males would pass the Army’s standard of no more than two children in addition
to a spouse, while only about 80 percent of Hispanic females would pass this standard. For
blacks, 85 percent of males would pass the Army standard compared to 71 percent of
females. The comparable figures for whites are 95 and 90 percent. Similar comparisons are
found for the other services.
Like their male counterparts, minority females face obstacles meeting the services’
AFQT standards. In the case of the Air Force, the fraction meeting the standard is 44
percent for Hispanic females and 38 percent for black females. For the Army, the figures are
78 percent and 68 percent, respectively. The figures are higher for the Army because of the
Army’s lower minimum score of 16. Still, the fraction who do not qualify for the Army on
the basis of AFQT represents a sizable proportion. As far as the other standards, drug use is
less important as an obstacle to enlistment for black and Hispanic females than it is for their
male counterparts.
Year One, Task Two: Assessing the Importance of Enlistment Standard Characteristics Among Hispanics
45
Hispanic Males Are More Likely Than
Blacks to Pass Standards
Cumulative Fraction Passing Standards for Diploma/GED, AFQT,
Weight, Dependents, Convictions, and Drugs (NLSY97)
Females
0.5
0.5
0.4
0.4
Fraction passing
Fraction passing
Males
0.3
0.2
0.1
0
0.2
0.1
0
White
Army
0.3
Air Force
Black
Navy
Hispanic
Marine Corps
Army
White
Black
Air Force
Navy
Hispanic
Marine Corps
-36
Cumulatively, accounting for all the standards using the NLSY data, we find that
among males, Hispanics are less likely to meet the standards than whites but more likely to
meet them than blacks. The cumulative fraction varies by service. The cumulative fraction
that meets the standards is lowest in the Air Force, and highest in the Marine Corps.
Among females, the cumulative fraction meeting all standards is about the same for
blacks and Hispanics, both being less than the fraction for their white counterparts. For
Hispanic females, the cumulative fraction is lowest for the Air Force and roughly equal for
the other three services.
CHAPTER FOUR
Year One, Task Three: Examining How the Fraction Passing
Enlistment Standards Has Changed over Time
Research Tasks
Year 1
• Task 1: Catalog current enlistment standards
• Task 2: Assess the importance of enlistment standard
characteristics among Hispanics
• Task 3: Examine how the fraction passing enlistment standards
has changed over time
Year 2
• Task 4: Estimate the representation of Hispanics in military
accession by Hispanic subgroup
• Task 5: Analyze the career and schooling paths of Hispanics
• Task 6: Analyze MEPCOM attrition among Hispanic applicants
• Task 7: Analyze the waiver process and the effects of enlistment
standards on the success of Hispanic recruits during their first
enlistment term
-37
The third project task was to consider the trend over time in the fraction of the
Hispanic youth population that would be eligible. We compare the results from the two
NLSY cohorts in 1997 and 1979 when survey respondents were ages 18 to 21 in 2001 and
1983, respectively. From the comparison, we can examine how qualification has changed
over the 20-year period, holding the standards at their current values. To make the AFQT
scores comparable across cohorts, we performed a principal components analysis to
transform the 1979 cohort’s raw ASVAB scores, just as we did for the 1997 cohort. The
derived AFQT index for each cohort indicates the relative position of individuals within the
AFQT distribution as measured in 1979 and in 1997.
47
48
What Factors Affect the Military Enlistment of Hispanic Youth?
Over Past 20 Years, More Hispanics Pass
Diploma and AFQT Standards; Fewer Pass
Weight Standards
Fraction of Males Passing Army Enlistment Standards
Fraction passing
1
0.8
0.6
0.4
0.2
0
Diploma/ AFQT Weight
GED
White
Diploma/ AFQT Weight
GED
Black
1980
Diploma/ AFQT
GED
Weight
Hispanic
2001
-38
This chart shows the results for the Army. We find a higher fraction of Hispanics and
blacks would pass the standard for education and AFQT now than they would 20 years ago,
but fewer would pass the Army’s weight standard. In the case of Hispanic males, the percent
of Hispanics ages 17 to 21 meeting the Army’s standard of a score of 16 on the AFQT was
57 percent in 1980 but was 71 percent in 2001 among the 1997 cohort. The improvement
was greater among blacks, rising from 47 percent to 65 percent. On the other hand, the
percent meeting the weight standard fell the most among blacks, followed by Hispanics, then
by whites.
Year One, Task Three: Examining How the Fraction Passing Enlistment Standards Has Changed over Time
49
Fewer Males Likely to Pass Drug Standards
Fraction of Males Without More Than Three Dependents, Conviction
of Crime, or Drug Use in Past Year
Fraction passing
1
0.8
0.6
0.4
0.2
0
Depen- Convic- Drugs
dents tions
Depen- Convic- Drugs
dents tions
White
Black
1980
Depen- Convic- Drugs
dents tions
Hispanic
2001
-39
This chart shows the trend between the 1979 and 1997 cohorts in the fraction that
would meet the Army’s standards in terms of having fewer than three dependents, no drug
use, or no convictions in the past year. Little change has occurred in the percent reporting
fewer than three dependents or no convictions. But the percent reporting no drug use has
fallen over time. For Hispanic males, the fraction who reported no drug use fell from 96 to
73 percent, while for white and black young men, it fell to 67 and 71 percent, from highs of
95 and 96 percent, respectively.
50
What Factors Affect the Military Enlistment of Hispanic Youth?
Fraction of male H.S. grads
with no...
Youth Poll Data Suggest Likelihood of
Passing Moral Standards Is Similar Across
Groups
1
0.8
0.6
0.4
0.2
0
White
Misdemeanor
Positive drug test
Black
Felony
Drug dependency
Hispanic
Judgment against
-40
The 2003 Youth Poll data provide more precise measures of moral character
disqualification standards, because the questions are more comparable to the definitions of
the standards than are the variables in the NLSY. The results are presented in this chart for
male high school graduates. About 80 percent of white males pass the standard of “no
misdemeanor conviction,” and nearly all graduates meet the “no felony conviction” standard,
as expected. The pattern for Hispanic young men is quite similar to that of white men and at
about the same magnitudes. Thus, Hispanic men do not seem to be less likely to meet these
standards than white or black males.
Year One, Task Three: Examining How the Fraction Passing Enlistment Standards Has Changed over Time
51
Summary of NLSY and Youth Poll Results
• AFQT is a significant obstacle to enlistment for
Hispanics, even after controlling for high school
diploma status
• Meeting the weight requirement is a growing problem
for Hispanics, especially for females
• Dependents, criminal activity, and drug use do not seem
to be significant obstacles to male enlistment
• Thus far, enlistment standards do not seem to explain
why Hispanics are so much less likely to enlist than
blacks
-41
To summarize the results, we find that AFQT is an important addition to the
traditional explanation of the underrepresentation of Hispanics among accessions, especially
for the Navy and Air Force. Even controlling for the lower high school graduation rates of
Hispanics, we find that they are less likely to meet the AFQT standards of the services than
their white counterparts for both males and females. Another important disqualifying factor,
especially for minority females, is weight. Among males, 28 percent of Hispanics and 29
percent of blacks would fail to meet the Army’s standards. The figures were 55 percent and
48 percent for females, respectively. In the case of weight, relaxing the standards would
improve eligibility more for Hispanic females. Our analysis did not suggest that dependent
status, criminal activity, or drug use were important disqualifying characteristics. However,
the NLSY variables were too broadly defined and could not account for waivers in the case of
criminal activity and drug use. Finally, our future analysis must reconcile our finding that
Hispanics are more likely to pass enlistment standards than are blacks, with the lower
enlistment rate among Hispanics. Since Hispanics have a higher enlistment propensity rate
than blacks, their lower enlistment rate cannot be due to their supply behavior.
52
What Factors Affect the Military Enlistment of Hispanic Youth?
What Can Be Done About AFQT and Weight?
• Implement DEP program aimed at improving language skills
•
•
•
•
– Relatively poor scores on AFQT for Hispanics could reflect
poor language ability
Implement programs to help applicants achieve weight standards
during DEP
– Implement programs to help with weight before entry
– Could mean higher attrition
Change weight standards to active duty standards
Stratify weight standards based on occupation
– Does not change fitness standards
– Provides incentives to exceed minimum standards
Increase recruiting in healthier populations
– Include college market
– Include recent immigrants
-42
What could be done to address weight problems of potential enlistees or to improve
the low AFQT scores of applicants? To the extent that low scores reflect English language
proficiency, the services could expand existing programs that target improvement in language
skills. In the case of weight, there are three alternative strategies.
First, the services might develop programs to help individuals reduce weight before
entry. For example, they might enroll individuals in specific weight-reduction programs
while in DEP. Whether such a program is cost effective would need to be evaluated. A
second alternative might be to relax the standards. The Army and Marines currently have
retention weight standards for active-duty personnel that differ from accession standards by
two to three pounds. Accession standards could be revised, and put in place for all branches,
to allow otherwise qualified individuals to enter and take off weight in DEP or boot camp
(basic training).
A third alternative would be to stratify the weight standards by what jobs individuals
perform. This acknowledges that different job functions require different sets of skills. Some
jobs are dependent on having a high AFQT score, yet all are required to meet the same
weight standard. Stratification of standards currently exists across branches. The Air Force
and Navy weight standards are not as stringent as the Army or Marine Corps standards, yet
personnel in these services can be just as effective at their service-specific mission. Such a
policy would not necessarily imply a change in fitness standards (e.g., the requirements for
running and doing sit-ups and push-ups could all be the same), and there are ways to provide
incentives to exceed the minimum standard. This policy would simply admit that not all
members are airborne infantrymen, and may be equally competent with different weight
requirements.
The final alternative is to recruit from healthier populations. The analysis of the
NHIS data indicated that those with more education, especially with some college and
Year One, Task Three: Examining How the Fraction Passing Enlistment Standards Has Changed over Time
53
noncitizens, are more likely to pass the weight standard. Enlistment programs that fast-track
the citizenship of immigrants serving in the military and recruitment programs targeting
college-bound youth may have the added benefits of improving the likelihood of youth
meeting weight standards or of improving Hispanic representation among enlistments.
Future research would need to investigate if this were the case.
54
What Factors Affect the Military Enlistment of Hispanic Youth?
Changing Weight or AFQT Standards Does
Not Change Eligibility Proportionately
Change in Fraction of Hispanic Males Eligible Is Due to Change in
Percentile Relative to Standard
Army
0.6
Fraction eligible
Fraction eligible
Air Force
0.6
0.5
0.4
0.3
0.2
0.1
0
-20
-10
10
20
0.1
0
-20
Weight
-10
0
10
AFQT
Weight
-20
-10
AFQT
0
10
Weight
20
Navy
Marine Corps
0.6
0.5
0.4
0.3
0.2
0.1
0
Fraction eligible
Fraction eligible
AFQT
0
0.5
0.4
0.3
0.2
20
0.6
0.5
0.4
0.3
0.2
0.1
0
-20
-10
AFQT
0
10
Weight
20
-43
It is important to recognize that policies that alter the standards for weight or AFQT,
say by defining job-specific weight standards or by granting waivers, would not necessarily
result in one-for-one increases or decreases in the fraction of the Hispanic population that
would be eligible for enlistment. The reason is that applicants may be disqualified based on
other characteristics that are correlated with AFQT score or weight, such as high school
graduation status or number of dependents.
These charts show how the fraction of Hispanic males who would be eligible for
enlistment would change as the weight standard or AFQT standard were tightened or
loosened relative to the current standard in each service. (Unlike in previous charts, we only
consider eligibility in terms of education, AFQT score, weight, and dependent status, and
ignore convictions and drug use.) For example, in the Army chart, about 43 percent of
Hispanic males would be eligible to enlist at current standards (indicated at 0 on the x axis).
If the weight standard were relaxed by 10 percentage points (to 10 on the x axis), the
percent eligible to enlist would increase by about four percentage points. Thus, relaxing the
weight standard so that 10 percent more individuals pass the weight standard would not
yield an increase in the fraction passing all enlistment standards by 10 percent, but only by
four percentage points. The fact that eligibility does not rise in direct proportion as the
weight standard relaxes means that other qualifying characteristics are negatively correlated
with weight. For example, many people who exceed the weight standard are not high school
graduates. Therefore, relaxing the weight standard in the Army will not improve eligibility
proportionately because these individuals will still be disqualified on the basis of their
education. For similar reasons, relaxing AFQT standards by 10 percentage points would
increase the fraction eligible in the Army chart to about 49 percent.
In Appendix C, we show similar diagrams for white and black males as well as for
black and Hispanic females. Three points are evident from these charts. First, the lines in the
Year One, Task Three: Examining How the Fraction Passing Enlistment Standards Has Changed over Time
55
figures are flat relative to the 45-degree lines, suggesting that correlations between the
qualifying characteristics are negative in general. Thus, in the case of Hispanic males,
relaxing the AFQT standard would not generate a proportionate increase in eligibility
because many individuals who have higher AFQT scores are not high school graduates.
Second, the slopes differ somewhat for AFQT versus weight, and the differences appear to
correlate with sex (see the backup charts for females). The slope of the eligibility line for
weight is steeper for Hispanic females than males but steeper for black males than females.
Thus, relaxing weight standards would improve eligibility more for Hispanic females and
black males. This result makes sense because females are more likely to be disqualified
because of weight. It also suggests that weight is less negatively correlated with other
qualifying characteristics among Hispanic females. As for AFQT, the slopes are steeper for
males than females for both blacks and Hispanics. Thus, relaxing AFQT standards would
improve male eligibility more than female eligibility among minorities.
Third, the slopes for Hispanics are considerably steeper than for whites for both
males and females. Thus, relaxing standards would improve eligibility rates more for
Hispanics than for whites. Stated differently, eligibility characteristics are less negatively
correlated for Hispanics than for whites in general.
CHAPTER FIVE
Year Two: Remaining Research Tasks
Research Tasks
Year 1
• Task 1: Catalog current enlistment standards
• Task 2: Assess the importance of enlistment standard
characteristics among Hispanics
• Task 3: Examine how the fraction passing enlistment
standards has changed over time
Year 2
• Task 4: Estimate the representation of Hispanics in
military accession by Hispanic subgroup
• Task 5: Analyze the career and schooling paths of
Hispanics
• Task 6: Analyze MEPCOM attrition among Hispanic
applicants
• Task 7: Analyze the waiver process and the effects of
enlistment standards on the success of Hispanic recruits
during their first enlistment term
-44
As stated at the outset, ours is a two-year project with seven tasks. The remaining
four tasks are building on the first-year results and focus on several topics. The unifying
theme is to provide information to develop policies to improve recruitment among the
Hispanic youth population. Task four is focusing on providing more detailed information on
Hispanic representation among accessions than what is currently available. For example, this
analysis is focusing on Hispanic subgroups, such as first- and second-generation immigrants.
Task five focuses on the career and school choices of Hispanic youth. An understanding of
these choices can provide insights on what policies may be developed to recruit Hispanic
youth more effectively. Task six is investigating how the services’ recruiting processes,
especially the applicant and the waiver processes, may affect Hispanic representation. Task
seven is investigating the implications of these processes for subsequent performance of
Hispanic recruits during their first term, where performance is measured by such proxy
variables as attrition rates and promotion rates to E-4. Our final report is forthcoming.
57
APPENDIX A
Revising Estimates of Hispanic Representation in the 18- to 24Year-Old Population
Appendix A: Revising Estimates of Hispanic
Representation in 18–24 Population
• Census weights are key ingredient to computing Hispanic
share of 18–24 civilian population
– Any analysis of civilian data, including the Current
Population Survey (CPS) or census data, must use
census weights to count the Hispanic population properly
• For example, Pop Rep figures are based on the CPS
• The 1990 Census undercounted the Hispanic population
– Major efforts were made to improve the 2000 Census
– Hispanic share of 18–24 population was undercounted
• Revised estimates of Hispanic share of 18–24 civilian
population in the 1990s are 2.2 percentage points higher
-45
Analyses of the social representation of the military proceed by comparing counts of
members of the military to counts of members of the corresponding civilian population. The
latter are drawn from DoD personnel files, while the former figures are traditionally drawn
from U.S. Census Bureau estimates. Recent technical changes in U.S. Census Bureau estimates are likely to have a substantial effect on the reported social representation of Hispanics
in the U.S. military. This appendix describes this issue in more detail. It begins by describing
the issue of weighting in general and then discusses the effects of weighting on civilian
population estimates.
For estimating representation, DoD uses estimates of the corresponding civilian
population derived from U.S. Census Bureau data. Specifically, footnotes in Pop Rep suggest
that DoD uses estimates derived from the counts of the population in the monthly Current
Population Survey (CPS) data. Thus, for the computation of most interest in this document,
DMDC takes the average of the weighted count of 18- to 24-year-old Hispanics over a given
fiscal year. We compute the weighted count of the 18- to 21-year-old Hispanic population
using the Demographic Supplement to the March CPS. The March survey questions refer to
59
60
What Factors Affect the Military Enlistment of Hispanic Youth?
outcomes in the previous year. However, demographic characteristics (e.g., age) and sample
weights are based on the survey month. Thus, for example, earnings reported in the March
1998 Demographic Supplement refer to earnings during calendar year 1997 (i.e., January
1997 to December 1997), but they are assigned to a respondent based on his or her age as of
March 1998 and based on the sample weights appropriate for March 1998. It is important,
however, to realize that the counts of the population from the CPS do not actually refer to
the information collected by the CPS. Instead, the U.S. Census Bureau implements a complicated set of procedures for adjusting the results of the CPS fieldwork. (For a complete discussion, see U.S. Census Bureau, 2002, Chapter 10, Chapter 11, and Appendix D, especially, pp. 10–15 and Table 10-2.) Here, we discuss the components of that process that are
relevant for the analysis presented below. We deliberately oversimplify or suppress many
other aspects of CPS procedures.
For our discussion, the crucial concept is a “control total.” For our purposes, the crucial control totals are the Census Bureau’s best estimates of the total civilian non-institutional
population for 14 Hispanic and five non-Hispanic age-sex categories. (Other control totals
consider the population by state, race, and finer age-sex categories.) They are derived, not
from the CPS, but from the presumed better population estimates from the decennial census
as adjusted for births and deaths from vital statistics data (i.e., birth and death certificates),
immigration (from a variety of sources), enrollment in the Armed Forces, and residence in
institutions.
Once a decade, after work on the decennial census is completed, the Census Bureau
revises its official control totals. As we will see below, this revision is sometimes significant,
causing major shifts in CPS-based estimates. This revision is usually implemented about
three years after the decennial census (i.e., in 1993 after the 1990 Decennial Census and in
2003 after the 2000 Decennial Census).
During the CPS fieldwork, random variation due to the CPS’s relatively small sample
size (only about 50,000 households) and systematic biases (e.g., inability to find or interview
certain types of individuals) lead to fractions of people in the sample that deviate from the
“known true” fractions from the control totals. To fix this problem, weights are created that
force the computed counts to conform exactly to the control totals for sex, age, and race and
ethnicity. Analysts then use these weights to compute counts and other estimates using the
CPS. That is, to ensure that the CPS reflects the U.S. population, analysts must adjust the
CPS data using the weights provided with the CPS data.
Of course, these weights and the resulting estimates will only be as good as the underlying control totals. In retrospect, it appears that the 1990 Decennial Census was of poor
quality (on the quality of the 1990 and 2000 Censuses, see Robinson, 2001). Large numbers
of minorities were “missed.” Furthermore, during the late 1980s and 1990s, there was large,
often undocumented, immigration. Both of these factors caused the control totals to underestimate the true numbers of Hispanics in the United States. The 2000 Census appears to
have been much more successful at counting minority individuals, including undocumented
immigrants currently living in the United States. Thus, estimates based on control totals using the 2000 Census are considered to be preferable to estimates based on control totals using the 1990 Census.
These technical issues have substantively large effects on estimates of the number of
Hispanics. Specifically, for CPS interviews in the early part of a decade, the census provides
weights either based on the control totals from the now-obsolete census (the 1990 Census) or
Revising Estimates of Hispanic Representation in the 18- to 24-Year-Old Population
61
based on the control totals computed based on the new census (the 2000 Census). Thus, we
can compute estimates for the same fieldwork, varying only the control totals (and the corresponding weights).
Our concept of interest is the share of Hispanics among 18- to 24-year-olds. Table
A.1 presents the underlying numbers. These figures are shown graphically in the main text of
the documented briefing. According to all measures, the fraction of Hispanics in the youth
population rose rapidly in the 1990s and early 2000s. This is due to a combination of immigration and births (the Hispanic population is relatively young and has relatively high agespecific fertility rates).
Our interest here, however, is on the effect of weighting. For each federal fiscal year
from 1993 to 2004, the table reports (approximately) comparable estimates of the fraction of
18- to 24-year-olds who are Hispanic. The first column gives the year to which the estimates
correspond. The second column gives the estimates from the March CPS of that year using
weights based on the 2000 Census. The third column gives the equivalent estimate based on
the 1990 Census. The fourth column reproduces Pop Rep’s published estimates (based on the
average of the estimates of the monthly CPS counts for October to September of that calendar year).
The final column, labeled “Adjusted,” gives an estimate for the entire period
1993–2004 that is consistent with the later estimates that use the 2000 Census-based
weights. When 2000 Census-based estimates are available, this adjusted figure uses that estimate. For the earlier period (1993–1999), this adjusted estimate uses a simple correction
based on the divergence between the 1990 Census- and 2000 Census-based estimates in
2000. In that year, the 2000 Census-based estimate is larger than the 1990 Census-based estimate by 2.2 percentage points. Consistent with this difference, the adjusted column adds
2.2 percentage points to each earlier year’s 1990 Census-based estimates.
Table A.1
Share of Hispanics Among 18- to 24-Year-Olds
Year
Weight 2000
Weight 1990
Pop Rep
Adjusted
1993
13.2
1994
13.3
13.0
15.5
1995
13.7
13.9
15.9
1996
14.5
14.3
16.6
1997
14.5
14.5
16.7
1998
15.0
15.0
17.2
1999
15.4
15.2
15.2
17.4
2000
17.1
14.9
15.0
17.1
2001
17.9
15.6
15.6
17.9
2002
17.9
17.9
2003
17.6
17.6
2004
17.9
17.9
SOURCES: Weight 2000: March CPS estimates using weights based on 2000 Census. Weight 1990: March CPS estimates using weights based on 1990 Census. Pop Rep: Estimates from Population Representation in the Military
based on average monthly CPS interview weights for the entire federal fiscal year (October through September).
NOTE: CPS interview occurred in March 2005.
62
What Factors Affect the Military Enlistment of Hispanic Youth?
Examination of Table A.1 and the figures earlier in this document show that the
March CPS estimates based on the 1990 Census weights and the Pop Rep estimates based on
the average of the monthly CPS weights for October through September are nearly identical.
Presumably any difference is due to using March rather than all 12 calendar months and
small differences between the March Supplement weight and the March monthly weight.
Comparisons of estimates based on the 1990 and 2000 Census weights are striking.
For the same year, the estimates based on the 2000 Census control totals are about two percentage points higher than those based on the 1990 control totals. These two percentage
points are about 25 percent of the 1990-based total. Recall that the 2000 Census is considered to have been of higher quality than the 1990 Census. Furthermore, even if the 1990
Census had been of the same quality as the 2000 Census, we would still prefer the 2000
Census-based control totals, because they are more recent. Thus, purely methodological considerations cause us to prefer the 2000 Census-based estimates for any years in which they
are available.
The Census Bureau has not backdated its control totals for earlier years in the 1990s.
However, a joint project of the National Cancer Institute and the National Center for
Health Statistics has done so. Ongoing work as part of this project will generate true adjusted
estimates of the Hispanic population for 1990–2000.
APPENDIX B
Enlistment Documents by Source
Appendix B: Enlistment Documents
by Source
Source
Document
DoD
DoD (1994).
Army
DoD (1995, 2002).
Air Force
DoD (2000).
Navy
U.S. Navy Recruiting Command (1979).
Marine Corps
DoD (1997).
-46
63
APPENDIX C
White and Black Males and Black, Hispanic, and White Females
Appendix C: White Males
Change in Fraction of White Males Eligible Is Due to Change in
Percentile Relative to Standard
Air Force
Fraction eligible
Fraction eligible
Army
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
-20
-10
0
10
20
AFQT
Weight
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
-20
Fraction eligible
Fraction eligible
-10
AFQT
0
10
Weight
0
10
20
Weight
Navy
Marine Corps
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
-20
-10
AFQT
20
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
-20
-10
AFQT
0
10
Weight
20
-47
This and the subsequent charts show the fraction of individuals in the NLSY sample
who would be eligible for enlistment as the weight standard or the AFQT standard were
tightened (moving right on each chart) or loosened (moving left) relative to the current
standard for each service. These charts correspond to the similar chart shown in the main
briefing for Hispanic males.
65
What Factors Affect the Military Enlistment of Hispanic Youth?
Appendix C: Black Males
Change in Fraction of Black Males Eligible Is Due to Change in
Percentile Relative to Standard
Air Force
0.6
0.5
Fraction eligible
Fraction eligible
Army
0.4
0.3
0.2
0.1
0
0.6
0.5
0.4
0.3
0.2
0.1
0
-20
-10
0
10
AFQT
Weight
20
-20
-20
-10
0
10
AFQT
Weight
0
10
20
Weight
Navy
Marine Corps
0.6
0.5
0.4
0.3
0.2
0.1
0
-10
AFQT
Fraction eligible
Fraction eligible
66
20
0.6
0.5
0.4
0.3
0.2
0.1
0
-20
-10
AFQT
0
10
Weight
20
-48
White and Black Males and Black, Hispanic, and White Females
Appendix C: Black Females
Change in Fraction of Black Females Eligible Is Due to Change in
Percentile Relative to Standard
Army
Fraction eligible
Fraction eligible
0.3
0.2
0.1
0.3
0.2
0.1
0
0
-10
0
10
AFQT
Weight
-20
20
-10
Marine Corps
0.2
0.1
0
-10
AFQT
0
10
Weight
10
20
Weight
Navy
0.3
-20
0
AFQT
Fraction eligible
Fraction eligible
-20
0.4
Air Force
0.4
0.4
20
0.4
0.3
0.2
0.1
0
-20
-10
AFQT
0
10
20
Weight
-49
67
What Factors Affect the Military Enlistment of Hispanic Youth?
Appendix C: Hispanic Females
Change in Fraction of Hispanic Females Eligible Is Due to Change in
Percentile Relative to Standard
Army
Air Force
0.5
Fraction eligible
Fraction eligible
0.5
0.4
0.3
0.2
0.1
0.4
0.3
0.2
0.1
0
0
-20
-10
0
10
AFQT
Weight
-20
20
Fraction eligible
0.5
0.4
0.3
0.2
0.1
0
-20
-10
AFQT
0
10
Weight
-10
AFQT
Marine Corps
Fraction eligible
68
20
0
10
20
Weight
Navy
0.5
0.4
0.3
0.2
0.1
0
-20
-10
AFQT
0
10
Weight
20
-50
White and Black Males and Black, Hispanic, and White Females
Appendix C: White Females
Change in Fraction of White Females Eligible Is Due to Change in
Percentile Relative to Standard
Air Force
Fraction eligible
Fraction eligible
Army
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
-20
-10
AFQT
0
10
Weight
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
20
-20
-10
AFQT
-20
-10
AFQT
0
10
Weight
20
Navy
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
Fraction eligible
Fraction eligible
Marine Corps
0
10
Weight
20
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
-20
-10
AFQT
0
10
Weight
20
-51
69
References
Burgos, Anthony E., Karen E. Schetzina, L. Beth Dixon, and Fernando S. Mendoza, “Importance of
Generational Status in Examining Access to and Utilization of Health Care Services by Mexican
American Children,” Pediatrics, Vol. 115, No. 3, 2005, p. 787.
Cho, Youngtae, and Robert A. Hummer, “Disability Status Differentials Across Fifteen Asian and
Pacific Islander Groups and the Effect of Nativity and Duration of Residence in the U.S.,” Social
Biology, Vol. 48, No. 3–4, 2001, pp. 171–195.
Defense Human Resources Activity, Department of Defense Youth Poll: Wave 5, Arlington, Va.: Defense Human Resources Activity, 2003.
DoD. See U.S. Department of Defense.
Franzini, Linda, John C. Ribble, and Arlene M. Keddie, “Understanding the Hispanic Paradox,”
Ethnicity and Disease, Vol. 11, 2001, pp. 496–518.
Frisbie, W. Parker, Youngtae Cho, and Robert A. Hummer, “Immigration and the Health of Asian
and Pacific Islander Adults in the United States,” American Journal of Epidemiology, Vol. 153, No.
4, 2001, pp. 372–380.
Hattiangadi, Anita U., Gary Lee, and Aline O. Quester, Recruiting Hispanics: The Marine Corps Experience Final Report, Alexandria, Va.: Center for Naval Analysis, CRM D0009071.A2/Final, January 2004. Online at http://www.cna.org/documents/D0009071.A2.pdf as of July 16, 2005.
Hoffman, Kathryn, Charmaine Llagas, and Thomas D. Snyder, Status and Trends in the Education of
Blacks, Washington, D.C.: National Center for Education Statistics, 2003. Online at http://purl.
access.gpo.gov/GPO/LPS38554 as of July 14, 2005.
Lindquist, Christine H., and Robert M. Bray, “Trends in Overweight and Physical Activity Among
U.S. Military Personnel, 1995–1998,” Preventive Medicine, Vol. 32, No. 1, 2001, pp. 57–65.
Llagas, Charmaine, and Thomas D. Snyder, Status and Trends in the Education of Hispanics, Washington, DC: National Center for Education Statistics, 2003. Online at http://purl.access.gpo.gov/
GPO/LPS32922 as of July 14, 2005.
Mendoza, Fernando S., and Elena Fuentes-Afflick, “Latino Children’s Health and the FamilyCommunity Health Promotion Model,” Western Journal of Medicine, Vol. 170, No. 2, 1999, pp.
85–92.
National Center for Health Statistics, Health, United States, 2003: With Chartbook on Trends in the
Health of Americans, Hyattsville, Md.: Department of Health and Human Services, 2003. Online
at http://www.cdc.gov/nchs/hus.htm as of July 14, 2005.
———, National Health Interview Survey , Hyattsville, Md.: National Center for Health Statistics,
1998. Online at http://www.cdc.gov/nchs/about/major/nhis/quest_data_related_1997_forward.
htm as of August 10, 2005.
71
72
What Factors Affect the Military Enlistment of Hispanic Youth?
———, National Health Interview Survey , Hyattsville, Md.: National Center for Health Statistics,
1999. Online at http://www.cdc.gov/nchs/about/major/nhis/quest_data_related_1997_forward.
htm as of August 10, 2005.
———, National Health Interview Survey , Hyattsville, Md.: National Center for Health Statistics,
2000. Online at http://www.cdc.gov/nchs/about/major/nhis/quest_data_related_1997_forward.
htm as of August 10, 2005.
———, National Health Interview Survey , Hyattsville, Md.: National Center for Health Statistics,
2001. Online at http://www.cdc.gov/nchs/about/major/nhis/quest_data_related_1997_forward.
htm as of August 10, 2005.
Nolte, Rochelle, Shawn C. Franckowiak, Carlos J. Crespo, and Ross E. Andersen, “U.S. Military
Weight Standards: What Percentage of U.S. Young Adults Meet the Current Standards?” American Journal of Medicine, Vol. 113, No. 6, 2002, pp. 486–490.
Palloni, Alberto, and Jeffrey D. Morenoff, “Interpreting the Paradoxical in the Hispanic Paradox:
Demographic and Epidemiologic Approaches,” Annals of the New York Academy of Sciences, Vol.
954, 2001, pp. 140–174.
Popkin, Barry M., and J. Richard Udry, “Adolescent Obesity Increases Significantly in Second and
Third Generation U.S. Immigrants: The National Longitudinal Study of Adolescent Health,” The
Journal of Nutrition, Vol. 128, No. 4, 1998, pp. 701–706.
Robinson, J. Gregory, ESCAP II: Demographic Analysis Results, Report No. 1, Washington, D.C.: U.S.
Census Bureau, 2001. Online at http://www.census.gov/dmd/www/pdf/Report1.PDF as of July
16, 2005.
Segall, Daniel, “Development and Evaluation of the 1997 ASVAB Score Scales,” Defense Manpower
Data Center, Draft V3.11.19, 2003.
U.S. Census Bureau, Census ’90, Washington, D.C.: U.S. Census Bureau, 1990. Online at http://
www.census.gov/main/www/cen1990.html as of July 12, 2005.
———, United States Census 2000, Washington, D.C.: U.S. Census Bureau, 2000. Online at http://
www.census.gov/main/www/cen2000.html as of July 12, 2005.
———, Current Population Survey: Design and Methodology, Washington, D.C.: U.S. Department of
Labor, Bureau of Labor Statistics, 2002. Online at http://purl.access.gpo.gov/GPO/LPS29233 as
of July 16, 2005.
———, Annual Demographic Survey: March Supplement, Washington, D.C.: U.S. Census Bureau
and U.S. Bureau of Labor Statistics, 2004. Online at http://www.bls.census.gov/cps/ads/sdata.
htm.
U.S. Department of Defense, Recruit Market Information System, undated. Online at https://www.
dmdc.osd.mil/rmis as of July 12, 2005.
———, Regular Army and Army Reserve Enlistment Program: Personnel Procurement, Washington,
D.C.: Department of the Army, Army Regulation 601-210, 1995.
———, Military Personnel Procurement Manual, Volume 2, Enlisted Procurement, Washington, D.C.:
U.S. Marine Corps, MCO P1100.72B, 1997.
———, Recruiting Procedures for the Air Force, Washington, D.C.: U.S. Air Force Education and
Training Command, AETCI36-2002, 2000.
———, Medical Services: Standards of Medical Fitness, Washington, D.C.: Department of the Army,
AR 40-501, 2002.
References
73
———, Population Representation in the Military Services Fiscal Year 2002 , Washington, D.C.: Office
of the Under Secretary of Defense for Personnel and Readiness, 2004. Online at http://www.
defenselink.mil/prhome/poprep2002 as of July 12, 2005.
U.S. Department of Labor, Bureau of Labor Statistics, National Longitudinal Surveys, National Longitudinal Survey of Youth 1979, Washington, D.C.: U.S. Department of Labor, 1979. Online at
http://www.bls.gov/nls/nlsy79.htm as of August 10, 2005.
———, National Longitudinal Survey of Youth 1997, Washington, D.C.: U.S. Department of Labor,
1997. Online at http://www.bls.gov/nls/nlsy97.htm as of August 10, 2005.
U.S. Navy Recruiting Command, Navy Recruiting Command, Enlisted, Arlington, Va.: Navy Recruiting Command, COMNAVCRUITROM INSTRUCTION 1130.8A, 1979.
Vega, W. A., and H. Amaro, “Latino Outlook: Good Health, Uncertain Prognosis,” Annual Review
of Public Health, Vol. 15, 1994, pp. 39–67.